Textkernel Parser July Release (2026.7)
What’s in this release
The July 2026 (2026.7) release for Textkernel Parser improves parsing accuracy for Hungarian-language CVs, address classification, and image-based documents, and resolves an issue with Indeed resume parsing. See the detailed release notes below for the full list of changes.
Improved Personal Information Parsing for Hungarian CVs
Extraction of names, email addresses, and phone numbers from Hungarian CVs has been improved, with personal information in Hungarian-language documents now identified more accurately during parsing.
UpdatedImproved Detection of Image-Based Files for OCR Processing
The parser now more reliably identifies image-based and scanned files and routes them through OCR processing before parsing, improving the accuracy and completeness of extracted candidate data.
UpdatedImproved Date Recognition in Hungarian CVs
Date recognition in the work experience and education history sections of Hungarian CVs has been improved, with a broader range of Hungarian date formats now correctly identified during parsing.
UpdatedImproved Country Classification in Address Parsing
Address parsing now correctly identifies the country across a broader range of address formats, improving the accuracy of location data extracted from CVs.
FixedParsing of Indeed Resumes Improved
Parsing of resumes sourced from Indeed has been improved following Indeed’s introduction of an AI Summary page. Previously, the AI Summary content affected how the rest of the resume was parsed, leading to inaccurate results.